Genetic Algorithm On Solving Job Shop Scheduling Problem

As a popular research realm in manufacture system, Job Shop Scheduling (JSP) is one of the most difficult problems in theoretics. The main task in scheduling, in terms of produce target and constraints, is to determine the precise process route, time, machine and operation et al for every process object. An eminent scheduling strategy can improve the optimization and economic efficiency of produce system.Due to the complexity such as constraints, nonlinearity, uncertainty and large scale of scheduling problem, many optimal algorithms like Simulant Anneal, Genetic Algorithm, Tabu Search, and Neural Network are developed. All these optimization algorithms propose new ways to solve complex problem; they work by simulating or disclosing natural phenomena, natural processes, and natural rules, where mathematics, physics, artificial intelligence et al are involved.So far, all the algorithms mentioned above arouse broad interests among scholars throughout the world and lifted research upsurge in this field because of their working mechanism, optimization capability and other particular advantages; they are applied successfully in many fields and solve many large quantities of difficult problems satisfactorily traditional optimization methods couldn't manage.In this paper, typical JSP problem is mainly introduced and its genetic algorithm is designed and realized. The coding issue and operation of algorithm in genetic algorithm based on JSP are given in detail; the "preference list-based representation" is annotated over again, and then the execution efficiency of this coding method is compared with other coding methods; In the end, a hybrid genetic algorithm PSA is designed, and the performance comparisons of PSA, OMSGA, and SA are presented.